This short article aims to help clinicians understand the implications of missing data due to dropout, a common problem that can affect the validity of clinical trial findings. Several analytic methods exist to handle this problem. Understanding the assumptions behind these methods can help clinicians decide whether and how to incorporate clinical trial findings into their own practice.
Use to:
- List reasons data might be missing in a clinical trial.
- Consider the example the authors provide of a trial assessing whether adding spinal manipulative therapy to home exercise is effective for back pain. How might each type of missing data affect the results?
- Review the definitions of missing at random, missing completely at random, and missing not at random. The authors note that it is important to understand what assumptions have been made about missing data. Why? How does it dictate the analytic approach? Invite an expert in epidemiology and/or biostatistics to join your discussion.
Annals of Internal Medicine is the premier internal medicine academic journal published by the Â鶹ֱ²¥app (ACP). It is one of the most widely cited and influential specialty medical journals in the world.
Back to May 2017 Issue of IMpact